search:

Perceptual ANalysis and DescriptiOn of Romanian visual Art


PN-II-RU-TE-2014-4-0733, contract 253 din 01.10.2015

Image Processing and Analysis Laboratory

Proiect finantat prin UEFISCDI

Pandora Database


We have collected from Internet and labelled two versions of the database:
  1. Pandora7k with 7740 images distributed among 12 classes. More details about the database and baseline results are to be found in the attached [report]

    Image Data is available as a .zip archive.
    Data for retrieving the reported results, is available as Matlab .mat files in this archive. Please note that images may be used for data mining/machine learning/retrival purposes



  2. Pandora18k with 18038 images distributed among 18 classes. The database was verified by engineers to ensure that only the relevant part of the images are shown and by an art expert to make sure that artistic movement annotation is true. More details about the database and baseline results are to be found [in this report] and [in this paper].

    Image Data is available as a .zip archive.
    Data for retrieving with standard image descriptors, is available as Matlab .mat files in this [archive].


Visual Saliency Data: 270 paintings from Pandora18k database have been annotated in terms of visual saliency. Details may to be found in this [paper]
Data (images, annotations, etc.) are to be found in this [archive].




Romanian Paintings Data: All paintings with romanian author from the WikiArt archive and other selected from Internet. Images are annotated with respect to style. Data (images, annotations, etc.) are to be found in this [archive].


Fair Use Notice


The datasets and this site contains material the use of which has not always been specifically authorized by the copyright owner. We do not own the copyright of the images. The material is made available in an effort to advance understanding of technological, scientific, and cultural issues. The material should be used for data mining/ information retrieval/ automatic image analysis purposes. We believe this constitutes a 'fair use' of any such material (according to paragraph i) from art 33. from law nr.8 /14.03.1996). The material on this site is distributed without profit to those who have expressed a prior interest in receiving the included information for non-commercial research and educational purposes. If you wish to use copyrighted material on this site or in the datasets for purposes of your own that go beyond non-commercial research and academic purposes, you must obtain permission directly from the copyright owner.




Back to Top

Artistic Photo Database

- scene annotations


Christopher Thomas and Adriana Kovashka for their paper ''Seeing Behind the Camera: Identifying the Authorship of a Photograph', CVPR 2016 collected a database of artistic photographs. The database can be retrieved their [project site].

We have experimented with domain transfer from artistic photographs to artistic paintings and for the process we have labelled a subset with respect to main subject (scene) of the photograph. The annotations are available in this [archive]. Annotations are in the following classes: Cityscape, Interior, Landscape, Portrait. To use them work, one needs to download the database and to write a code that parse the list so to select from the database the images with annotations.

The results obtained in this experimentation may be followed in the following [report]
.
Back to Top